Credit Default Swaps vs. InsuranceMedium

Both pay out when something goes wrong. Only one requires you to own the thing, and that single difference decides who may buy it, how it is priced, who regulates it and what happens in a crisis.

5 min read · 888 words

The analogy, and where it stops

• The similarity is real. A buyer pays a periodic amount and receives a payment if a defined bad thing happens to a named party. Described at that altitude the two are the same product.
• The first difference decides all the others. An insurance contract requires insurable interest: you may insure a house you own, and not one you merely have views about. A credit default swap requires nothing of the kind — you may buy protection on a borrower you have never lent a euro to.
• That is not a loophole; it is the design. Without it there is no market, because the whole point of a traded credit instrument is that anybody may take either side. With it, the instrument stops being a hedge and becomes a position — and once it can be a position, it can be a very large one relative to the debt it references.
• So "it is basically insurance" is a useful first sentence and a bad second one, and this page is about the second one. The CDS page covers the instrument; this is what it is not.

Indemnity against a defined payout

• Insurance indemnifies. It restores you to where you were, and no further: the claim is assessed against the actual loss suffered. You cannot profit from an insured event, and a policy that let you would be void.
• A credit default swap pays a formula. On a credit event, protection pays par less the recovery value determined at an auction, whatever the buyer's own position happens to be. A buyer with no exposure receives the same as one who is fully exposed.
• The auction is the mechanism that makes that possible. One market-wide determination of what the defaulted debt is worth, so every contract settles against the same number rather than against each holder's circumstances.
• Which is why the outstanding contracts can exceed the debt. Nothing in the instrument ties the number of contracts to the amount borrowed, and net exposure after offsetting positions is a different and much smaller figure than gross — a distinction getting it right in writing names as one of the four errors that survive every edit.

Who decides that the bad thing happened

• An insurer assesses the claim, and the policyholder may dispute it. The insurer is a party to the contract and to the decision, which is a conflict the whole of insurance regulation is built around.
• A credit event is determined by a committee convened for the market as a whole, against defined triggers — failure to pay, bankruptcy, restructuring — rather than case by case. Neither party to a given contract decides.
• Restructuring is the trigger that keeps being argued about, because a negotiated change to a borrower's terms may or may not be the thing protection was bought against. The Greek restructuring of 2012 is where that argument ran in public, and it is why the definitions have been rewritten more than once.

What stands behind the promise

• An insurer reserves. It holds assets against expected claims, is capitalised for the unexpected ones, and is supervised as an insurer — with rules about what it may hold and how it must value what it owes.
• A protection seller posts collateral. Standard contracts are collateralised daily against the mark, and increasingly cleared, which replaces "the seller will be good for it" with "the seller has already posted most of it".
• The difference is when the money arrives. An insurer's reserve is there because a claim might come; a swap's collateral moves as the mark moves, every day, which is a funding obligation rather than a provision. Margin and collateral covers what that does to a seller.
• Uncollateralised protection is where this broke. A seller writing large volumes without posting against the mark has sold insurance without reserving for it, and the exposure appears only when the marks move — which is the structure at the centre of 2008.

The comparison

What the comparison is actually good for

• Explaining the payoff to somebody new — the insurance analogy does that job in one sentence and nothing else does it as well.
• And then immediately stopping. Every question that follows the first one — who may buy it, what it pays, who decides, what backs it, how much can exist — has a different answer in the two, and the answers are not close.
• The general lesson is the one what the wrapper changes makes across the site: two contracts with a similar payoff can differ in every respect that decides what happens when the payoff is due.